Add HEALTHCHECK instruction to Dockerfile #1546

New issue

Open

opened 2026-03-16 14:57:20 +00:00 by s1lv3r · 5 comments

s1lv3r commented

2026-03-16 14:57:20 +00:00

Contributor

For readiness monitoring; see https://matrix.to/#/!ksTlboXVgcyWjv5GrlEeKyQuJ8ZCprnwQx2b6-BQ44Q/%244PvqwxuCcNPYTota6PiTHO8ORSq2V5iWS6WUD39fNEY?via=wolfgirl.pet&via=gingershaped.computer&via=continuwuity.org and below

s1lv3r commented

2026-03-16 15:09:11 +00:00

Author

Contributor

I was just reminded why this is more challenging than first expected: The docker image contains no way of running a web request, therefore something simple like curl won't work. We'd have to add a way to perform web requests, possibly a static busybox?

Jade commented

2026-03-16 16:24:31 +00:00

Owner

Can maybe see if pulling in a curl from the host image works, seeing as we already bring in glibc? Alternatively adding a basic http request command in the main binary is what I've done in another project 🥴

s1lv3r commented

2026-03-16 16:26:24 +00:00

Author

Contributor

That is so utterly cursed.

Pulling in curl from the host image could work, however then we'd have to pull in all its deps and stuff as well since its dynamically linked. Would be a lot more complicated than a static busybox

That is so utterly cursed. Pulling in curl from the host image _could_ work, however then we'd have to pull in all its deps and stuff as well since its dynamically linked. Would be a lot more complicated than a static busybox

Jade commented

2026-03-16 16:40:53 +00:00

Owner

Sounds like not a bad idea, although it would mean pulling in yet more images at build time. Up to whoever does it though

stavros-k commented

2026-03-16 18:14:07 +00:00

A subcommand to the continuwuity would solve this issue.
ie
/sbin/continuwuity healthcheck

This can load the env vars and do the request to the healthz endpoint, exiting with 0 on success and 1 on failure.

A subcommand to the continuwuity would solve this issue. ie /sbin/continuwuity healthcheck This can load the env vars and do the request to the healthz endpoint, exiting with 0 on success and 1 on failure.